Bruce A. Craig

Written by: Shannon Knapp, Ph.D. candidate in Statistics 


Accurately monitoring the population of wildlife is a task that lends itself well to statistical analysis. Professor Bruce Craig has been working with Dr. John Reynolds, Chairman of the Marine Mammal Commission and Manager, Manatee Research Program, Mote Marine Laboratory in Sarasota, Florida, to employ statistical methods to analyze years of collected data on the federally endangered Florida manatee.

Manatees, also known as sea cows, are fully aquatic marine mammals. They survive by grazing on sea grass and other aquatic vegetation along the coast of the southeastern U.S., primarily concentrated along the coast of Florida. The growing human population in Florida and the impact of many human activities such as boating and land development in the manatees' habitat threaten the population. In order to make rational decisions about the management of the manatee population, good scientific studies are essential. 

Aerial surveys are one way in which the manatee population is monitored. Because manatees are unable to survive for long periods in water below 68 degrees Fahrenheit, they will congregate in warm water, including naturally occurring springs and coastal power plants that discharge warm water, when weather conditions become unfavorable. Since 1977 Reynolds has taken advantage of these aggregations, flying over the sites and counting the manatees. However, while a simple count seems straight-forward, there are many reasons for counts to vary from year to year. Aside from changes in the population size, sighting conditions can vary. For example, wind can make the water choppy and the sun can cause surface glare, making it difficult to count the manatees. In addition, the severity of a cold front can affect how likely it is that a manatee will seek out one of these warm water sites. Finally, manatees are able to move between sites between years or even between surveys taken within a single year.

Aerial View of Manatees

Craig worked with Reynolds and other manatee experts to develop a Bayesian hierarchical model to describe these counts.  He states, "there is a tremendous amount of variability in these data and hierarchical modeling provides a straightforward way to synthesize known sources of variation.  This model accounts for differences in sighting quality among surveys, movement of the manatees, and most importantly, changes in the number of manatees from year-to-year."   Craig's analysis shows that the manatee populations along the Atlantic coast of Florida have been relatively stable, and have even been increasing at a rate of 3-4% per year. "Earlier analyses of these data focused on linear regression models of an overall survey count and could only address the rate of change.  Our model not only handles the relationship among site counts within a survey but also allows us to estimate the total population size." (Craig, B.A. and Reynolds III, J.E. (2004). Determination of manatee population trends using a Bayesian approach with temperature-adjusted aerial survey data, Marine Mammal Science, 20, 386-400.)

Craig joined the Purdue University Department of Statistics in 1996. He is one of the leaders in the Department's interdisciplinary efforts. His research focuses primarily on the utilization of hierarchical models to address problems in the life sciences. Application areas of his research include diagnostic testing, protein structure determination, plant and human genetics, wildlife population demographics, disease progression modeling, and medical decision making. Craig began working with the Statistical Consulting Center in 2000, and became director of the Statistical Consulting Center in 2004. For more information about Professor Craig's research, please visit his home page

Last Updated: Sep 11, 2017 1:58 PM

Purdue Department of Statistics, 250 N. University St, West Lafayette, IN 47907

Phone: (765) 494-6030, Fax: (765) 494-0558

© 2015 Purdue University | An equal access/equal opportunity university | Copyright Complaints

Trouble with this page? Disability-related accessibility issue? Please contact the College of Science Webmaster.